US6236680B1 - Encoding and decoding system of motion image containing arbitrary object - Google Patents

Encoding and decoding system of motion image containing arbitrary object Download PDF

Info

Publication number
US6236680B1
US6236680B1 US09/504,934 US50493400A US6236680B1 US 6236680 B1 US6236680 B1 US 6236680B1 US 50493400 A US50493400 A US 50493400A US 6236680 B1 US6236680 B1 US 6236680B1
Authority
US
United States
Prior art keywords
data
image
mesh
motion
predictive
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Lifetime
Application number
US09/504,934
Inventor
Kang-wook Chun
Byeung-woo Jeon
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Samsung Electronics Co Ltd
Original Assignee
Samsung Electronics Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Samsung Electronics Co Ltd filed Critical Samsung Electronics Co Ltd
Priority to US09/504,934 priority Critical patent/US6236680B1/en
Priority to US09/842,405 priority patent/US6744817B2/en
Application granted granted Critical
Publication of US6236680B1 publication Critical patent/US6236680B1/en
Anticipated expiration legal-status Critical
Expired - Lifetime legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
    • H04N19/51—Motion estimation or motion compensation
    • H04N19/537—Motion estimation other than block-based
    • H04N19/54—Motion estimation other than block-based using feature points or meshes
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/42—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation
    • H04N19/43—Hardware specially adapted for motion estimation or compensation
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
    • H04N19/51—Motion estimation or motion compensation
    • H04N19/537—Motion estimation other than block-based
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
    • H04N19/61—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding in combination with predictive coding
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/20—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using video object coding

Definitions

  • the present invention relates to an encoding and decoding system of a motion image containing an arbitrary object.
  • an object-oriented encoding technique for partitioning a meaningful object in the motion image and transmitting the partitioned object is being widely studied.
  • the object-oriented encoding technique is recognized as an essential technique in order to transmit a motion image at a very low bit rate.
  • a motion prediction technique for removing temporal correlation with respect to an object is required and a more accurate motion prediction technique is also needed for enhancing encoding efficiency.
  • Another object of the present invention is to provide a decoding apparatus for decoding data encoded in the above encoding apparatus.
  • a motion image encoding apparatus for encoding a current image containing an object
  • the motion image encoding apparatus comprising: an object extraction unit for extracting the object contained in the current image from received current image data and outputting object contour data indicating a contour of the extracted object; a predictive encoding unit for performing a predictive encoding operation using the current image data, prestored data of a reference image, and control points of meshes which divide one of the current image and the reference image, and generating motion information involved with the control points, differential data between the current image and the reference image, and predictive image data; an object difference generator for selecting differential data in an object region among the differential data supplied from the predictive encoding unit to encode the selected differential data based on the object contour data output from the object extraction unit, and generating the selected differential data and the encoded differential data; and an addition unit for receiving and adding the predictive image data output from the predictive encoding unit and the differential data selected in the object difference generator, and updating the reference image data prestored
  • a motion image decoding apparatus for decoding the output of a motion image encoding apparatus, the motion image decoding apparatus comprising: an object contour restoring unit for decoding received encoded object contour data and restoring object contour data indicating a contour of an object contained in a current image; a differential data decoding unit for decoding the received encoded differential data and outputting differential data; a motion compensation decoding unit for performing a motion compensation operation using the object contour data, prestored reference image data, control points of meshes which divide one of an object contour and a reference image, and motion information involved with the control points; and an addition unit for adding the predictive image data selected by the object contour data and the differential data, and updating the reference image data prestored in the motion compensation decoding unit using the image data obtained by the addition result.
  • FIG. 1 is a block diagram showing a motion image encoding apparatus according to a preferred embodiment of the present invention
  • FIG. 2 is a detailed block diagram of the mesh producer of FIG. 1;
  • FIG. 3 is a block diagram showing a motion image decoding apparatus for decoding data encoded by the FIG. 1 apparatus;
  • FIG. 4 is a block diagram showing a motion image encoding apparatus according to another preferred embodiment of the present invention.
  • FIG. 5 is a block diagram showing a motion image decoding apparatus for decoding data encoded by the FIG. 4 apparatus.
  • an object extractor 10 receives digital image data from an external source, and extracts an arbitrary object from a current image represented by the received image data. Then, the object extractor 10 generates object contour data representing contour information of the extracted object.
  • the object contour data is supplied to a differential data encoder 20 and a decoding apparatus of FIG. 3 to be described later.
  • the object contour data contains spatial positions of pixels which determine a contour of the object or spatial positions of segments composed of the pixels.
  • the object extractor 10 outputs the object contour data and the received current image data to a mesh generator 12 .
  • the mesh generator 12 discriminates an object region from the current image based on the object contour data, and performs a signal processing for regular mesh representation with respect to the image in the object region. The detailed structure and operation of the mesh generator 12 will be described with reference to FIG. 2 .
  • a block former 121 in the mesh generator 12 receives the current image data and the object contour data output from the object extractor 10 .
  • the block former 121 divides the current image represented by the received image data into image blocks each having a predetermined size, and supplies image block data representing image blocks and the object contour data to a block selector 123 .
  • the block selector 123 selects image blocks containing the image data in the object region among the entire image blocks constituting the current image based on the object contour data.
  • the data representing the selected image blocks and the object contour data is output to an object image composer 125 .
  • image blocks having only image data which do not belong to the object region are not selected. Therefore, the data representing the image blocks which are not selected is not supplied to the object image composer 125 .
  • the object image composer 125 which receives the data of the selected image blocks merges the selected image blocks and supplies the data representing the object image generated from the merging result and the object contour data to a mesh composer 127 .
  • the mesh composer 127 receiving the object image data divides the object image into a regular rectangular mesh or a regular triangular mesh.
  • the mesh composer 127 divides the object image into grids each having a predetermined size different from a block size.
  • a grid has a size smaller than a block.
  • the mesh composer 127 determines vertices contained in the object region among the vertices of the grids obtained by the division as control points, and determines grids having the control points as regular rectangular meshes.
  • the mesh composer 127 outputs rectangular mesh data containing the image data involving the control points of the regular rectangular meshes to a motion estimation and compensation portion 14 of FIG. 1 .
  • the image data involving each control point includes the position of a control point and a gray value at the position of the control point, and the positions of the pixels having positions adjacent to the control point and the gray values.
  • the mesh composer 127 also outputs the current image data to a difference value calculator 18 .
  • the mesh composer 127 divides each regular rectangular mesh obtained through the above process into two regular triangular meshes based on similarity of the image data in the mesh. To judge similarity of the image data of each regular rectangular mesh, the mesh composer 127 compares a difference value between pixel values corresponding to two control points which are located at a diagonal direction of 45° in the regular rectangular mesh, with a difference value between pixel values corresponding to two control points located at a diagonal direction of ⁇ 45° therein in size. The mesh composer 127 divides the rectangular mesh in the diagonal direction corresponding to the smaller difference value. As a result, two regular triangular meshes are generated for each rectangular mesh. The mesh composer 127 outputs the triangular mesh data containing the image data involving the control points of the regular triangular meshes to the motion estimation and compensation portion 14 . The mesh composer 127 outputs the current image data to a difference value calculator 18 as well.
  • the motion estimation and compensation portion 14 performs motion estimation and compensation using the reference image data stored in the memory 16 and the mesh data involving the object image in the current image received from the mesh former 12 .
  • the motion estimation and compensation portion 14 first compares in magnitude the image data involved in the control points of each mesh with the reference image data stored in the memory 16 .
  • Each control point ideally corresponds to a single pixel. That is, a control point is expressed by a spatial position of a pixel and a gray value. However, it is actually difficult to exactly find out a position having image data similar to the control point in the reference image by only pixel to pixel comparison.
  • the image data involving each control point is image data which is involved in a pixel corresponding to the control point and a plurality of pixels neighboring the pixel corresponding to the control point, as described in the mesh composer 127 .
  • the motion estimation and compensation portion 14 determines a position in the reference image having the image data which is most similar to the image data of the control point based on the data magnitude comparison result. Then, the motion estimation and compensation portion 14 determines a motion vector representing movement between a control point and a corresponding position in the reference image. If the motion vectors are determined with respect to all the control points, the motion estimation and compensation portion 14 performs spatial transformation such as image warping using the determined motion vectors, and generates a predictive image with respect to the current image.
  • the data representing the predictive image is output to the difference value calculator 18 and the adder 24 .
  • the motion information representing the motion vectors corresponding to all the control points is supplied to the FIG. 3 apparatus. Since the above-described spatial transformation is well known in the field of a digital image processing, the detailed description thereof will be omitted.
  • the difference value calculator 18 calculates difference values between the current image data received from the mesh generator 12 and the predictive image data output from the motion estimation and compensation portion 14 .
  • the difference value calculation is performed between the pixels having corresponding positions with respect to the current image and the predictive image, and the resultant difference value data is output to a differential data encoder 20 .
  • the differential data encoder 20 encodes difference value data in the object region determined by the object contour data of the object extractor 10 among the difference value data supplied from the difference value calculator 18 . If the data encoded by the differential data encoder 20 is defined as difference value data in the object region, only difference value data involving the object of the current image is encoded. Accordingly, the differential data encoder 20 can perform a more efficient encoding operation with respect to the current image.
  • the differential data encoder 20 encodes the object contour data received from the object extractor 10 as well.
  • the differential data encoder 20 uses a well-known orthogonal transform coding method including discrete cosine transform (DCT) and so on.
  • DCT discrete cosine transform
  • the encoded object contour data, the encoded difference value data and the motion information are transmitted to a decoding apparatus of FIG. 3 to be described later via a transmission channel (not shown), or are recorded on a storage medium used in the decoding apparatus of FIG. 3 .
  • a differential data decoder 22 receiving the encoded difference value data from the differential data encoder 20 restores the difference value data via a reverse procedure of the signal processing of the differential data encoder 20 .
  • the adder 24 adds the predictive image data output from the motion estimation and compensation portion 14 and the difference value data of the differential data decoder 22 , and outputs the added result to the memory 16 .
  • the data output from the adder 24 is data of the current image from which motion estimation and compensation has been performed, and is stored in the memory 16 to be used as reference image data for motion estimation and motion compensation with respect to a next image.
  • a motion image decoding apparatus shown in FIG. 3 receives the encoded difference value data, the encoded object contour data and the motion information generated by the FIG. 1 apparatus.
  • An object contour restorer 30 decodes the encoded object contour data.
  • the object contour data is output to a mesh generator 34 .
  • the mesh generator 34 receiving the object contour data divides the entire image into grids each having a predetermined size, determines vertices existing in the object region determined by the object contour data as control points among the vertices of the grids obtained by the division, and determines the grids having the control points as rectangular meshes.
  • the mesh generator 12 of FIG. 1 is designed to generate mesh data representing the rectangular meshes
  • the mesh generator 34 also generates mesh data representing rectangular meshes.
  • the mesh generator 34 generates mesh data involving the triangular meshes.
  • the mesh generator 34 supplies the generated mesh data to a motion compensator 38 and supplies the object contour data to an adder 36 .
  • a differential data decoder 32 decodes the encoded difference value data generated by the FIG. 1 apparatus.
  • the difference value data is output to the adder 36 .
  • the motion compensator 38 receives the mesh data output from the mesh generator 34 and the motion information generated by the FIG. 1 apparatus.
  • the motion compensator 38 generates predictive image data using control points contained in the mesh data corresponding to a current image, motion vectors corresponding to all the control points and contained in motion information, and the reference image data stored in the memory 40 .
  • the motion compensator 38 uses a motion vector corresponding to each control point to find out a position in the reference image which is most similar to the control point, and then uses spatial transformation such as image warping to generate predictive image data for a current image.
  • the motion compensator 38 outputs the generated predictive image data to an adder 36 .
  • the adder 36 receives the predictive image data output from the motion compensator 38 and the difference value data output from the differential data decoder 32 .
  • the adder 36 adds only predictive image data in the object region determined by the object contour data supplied from the mesh generator 34 among the entire predictive image data to the difference value data corresponding thereto. By doing so, image data with respect to an object in the current image can be restored.
  • the output data of the adder 36 is stored in the memory 40 to be used as reference image data for motion compensation of a next image.
  • FIG. 4 is a block diagram showing a motion image encoding apparatus according to another preferred embodiment of the present invention. Since the blocks of FIG. 4 having the same reference numerals as those of the FIG. 1 blocks perform the same functions as those of the corresponding blocks of FIG. 1, the detailed description thereof will be omitted.
  • the FIG. 1 apparatus uses the meshes obtained from the current image and the reference image data stored in the memory 16 to generate predictive image data. On the contrary, the FIG. 4 apparatus uses meshes obtained from reference image data stored in a memory 48 and externally input current image data to generate predictive image data.
  • the external current image data is input to the object extractor 10 and a motion estimation and compensation portion 44 .
  • the object extractor 10 generates object contour data using the current image data.
  • the object contour data is output to the differential data encoder 20 .
  • a mesh generator 42 reads the reference image data stored in the memory 48 and divides the entire reference image into irregular meshes. For such division, the mesh generator 42 divides the reference image into regular rectangular meshes and divides the regular rectangular meshes into regular triangular meshes. Since the method for generation of the regular rectangular meshes and the regular triangular meshes is the same as a signal processing of the FIG. 1 mesh generator 12 , the detailed description thereof will be omitted.
  • Another function of the mesh generator 42 generates an irregular mesh from regular rectangular meshes or regular triangular meshes.
  • the procedure of generating the irregular mesh is as follows.
  • the mesh generator 42 removes one selected among the control points of the regular triangular meshes, and performs triangulation with respect to the region from which the control point is removed.
  • the triangular meshes generated by the triangulation have irregular triangular shapes.
  • Such control point removal and triangulation repeats until the number of the remaining control points is identical to a predetermined value.
  • a quantified difference value is obtained between image descriptiveness obtained when the control point has been removed from a support region of the control point and that when the former has not been removed.
  • the support region is a region surrounded by neighboring control points to a corresponding control point and lines connecting the neighboring control points.
  • a control point corresponding to the smallest difference value among the quantified difference values of the image descriptiveness contributes little to the image descriptiveness, and accordingly can be removed.
  • the irregular mesh generation technique is disclosed in a paper entitled “Irregular Triangular Mesh Representation Based on Adaptive Control Point Removal” published in SPIE's 1996 Symposium on Visual Communications and Image Processing by Kang W. Chun, Byungwoo Jean and Jae M. Jo.
  • the mesh generator 42 outputs information about the finally generated irregular triangular meshes, that is, the irregular mesh data representing the remaining control points and the irregular meshes surrounded by the control points, to a motion estimation and compensation portion 44 .
  • the mesh generator 42 outputs the reference image data read from the memory 48 to a difference value calculator 46 as well.
  • the motion estimation and compensation portion 44 receives the externally supplied current image data and the irregular mesh data of the mesh generator 42 , and reads the reference image data stored in the memory 48 .
  • the motion estimation and compensation portion 44 determines corresponding positions in the current image having the image data which is most similar to the control points of the irregular meshes. After the corresponding positions are determined, the motion estimation and compensation portion 44 generates motion vectors between the control points and the corresponding positions and uses spatial transformation to generate a predictive image using the current image.
  • the predictive image data is supplied to the difference value calculator 46 and the adder 24 , and the motion information representing the motion vectors are transmitted via a transmission channel or recorded on a storage medium to be used for a decoding apparatus.
  • the difference value calculator 46 generates difference value data between the reference image data received from the mesh generator 42 and the predictive image data from the motion estimation and compensation portion 44 , to supply the difference value data to a differential data encoder 20 .
  • the differential data encoder 20 encodes the object contour data output from the object extractor 10 and the difference value data output from the difference value calculator 46 .
  • the differential data encoder 20 processes the input data in the same manner as those of the corresponding blocks of FIG. 1 .
  • the encoded data output from the differential data encoder 20 includes data obtained by encoding the difference value data in the object region and the encoded object contour data.
  • the encoded data is transmitted to a receiver via a transmission channel (not shown), or is recorded on a recording medium (not shown).
  • the encoded difference value data among the encoded data is transmitted to the differential data decoder 22 .
  • the differential data decoder 22 decodes the difference value data encoded in the differential data encoder 20 , to supply the decoded difference value data to the adder 24 .
  • the adder 24 adds the predictive image data supplied from the motion estimation and compensation portion 44 and the difference value data of the differential data decoder 22 , and the resultant data is stored in the memory 48 , to be used as reference image data for motion estimation and motion compensation for a next image.
  • FIG. 5 shows a motion image decoding apparatus corresponding to the FIG. 4 apparatus. Since blocks having the same reference numerals as those of the FIG. 3 blocks among the blocks shown in FIG. 5 have the same functions as those of the FIG. 3 blocks, the detailed description thereof will be omitted.
  • the encoded object contour data is input to an object contour restorer 30 and the motion information is input to a motion compensator 52 . Also, the encoded difference value data is supplied to the differential data decoder 32 .
  • the differential data decoder 32 decodes the encoded difference value data to output the decoded difference value data to an adder 56 .
  • the object contour restorer 30 decodes the encoded object contour data to output the decoded object contour data to a mesh former 51 .
  • the mesh former 51 reads the reference image data stored in a memory 54 , and divides the object region in the reference image determined by the object contour data via the same signal processing as that of the FIG. 4 mesh generator 42 , into irregular meshes.
  • the mesh data output from the mesh former 51 is supplied to the motion compensator 52 .
  • the motion compensator 52 uses the mesh data, the motion information and the reference image data stored in the memory 54 to generate the predictive image data.
  • the adder 56 adds the predictive image data and the difference value data output from the differential data decoder 32 .
  • the resultant current image is stored in the memory 54 to be used as reference image data for motion compensation of a next image.
  • the encoding and decoding system performs motion estimation and motion compensation based on mesh representation with respect to an arbitrary object contained in an image. As a result, a more efficient encoding can be performed compared with a case when motion estimation and motion compensation is performed with respect to the entire image. Thus, an apparatus appropriate for applications requiring a very low bit rate can be designed.

Abstract

A motion image encoding apparatus for a current image containing an object uses mesh representation for encoding a motion image. The encoding apparatus includes an object extractor for extracting the object contained in the current image from received current image data and outputting object contour data indicating a contour of the extracted object. A predictive encoder performs a predictive encoding operation using the current image data, prestored data of a reference image, and control points of meshes which divide one of the current image and the reference image, and generates motion information relating to the control points, differential data between the current image and the reference image, and predictive image data. An object difference generator selects differential data in an object region among the differential data supplied from the predictive encoder to encode the selected differential data based on the object contour data output from the object extractor, and generates the selected differential data and the encoded differential data. An adder receives and adds the predictive image data output from the predictive encoder and the differential data selected in the object difference generator, and updates the reference image data prestored in the predictive encoder using the image data obtained by the addition result.

Description

This application is a continuation of Ser. No. 08/864,992 filed May 29, 1997, U.S. Pat. No. 6,038,258.
BACKGROUND OF THE INVENTION
The present invention relates to an encoding and decoding system of a motion image containing an arbitrary object.
Recently, international standards for transmitting a motion image at a very low bit rate have been progressively advancing in order to be applied to a video phone or a video conference system. To transmit a motion image at a very low rate, an object-oriented encoding technique for partitioning a meaningful object in the motion image and transmitting the partitioned object is being widely studied. The object-oriented encoding technique is recognized as an essential technique in order to transmit a motion image at a very low bit rate. In the object-oriented encoding of a motion image, a motion prediction technique for removing temporal correlation with respect to an object is required and a more accurate motion prediction technique is also needed for enhancing encoding efficiency.
SUMMARY OF THE INVENTION
It is an object of the present invention to provide an encoding apparatus for a motion image containing an arbitrary object in which the arbitrary object extracted from an image is represented as meshes, and control points of the meshes are used for motion predictive encoding of the image.
Another object of the present invention is to provide a decoding apparatus for decoding data encoded in the above encoding apparatus.
To accomplish the above object of the present invention, there is provided a motion image encoding apparatus for encoding a current image containing an object, the motion image encoding apparatus comprising: an object extraction unit for extracting the object contained in the current image from received current image data and outputting object contour data indicating a contour of the extracted object; a predictive encoding unit for performing a predictive encoding operation using the current image data, prestored data of a reference image, and control points of meshes which divide one of the current image and the reference image, and generating motion information involved with the control points, differential data between the current image and the reference image, and predictive image data; an object difference generator for selecting differential data in an object region among the differential data supplied from the predictive encoding unit to encode the selected differential data based on the object contour data output from the object extraction unit, and generating the selected differential data and the encoded differential data; and an addition unit for receiving and adding the predictive image data output from the predictive encoding unit and the differential data selected in the object difference generator, and updating the reference image data prestored in the predictive encoding unit using the image data obtained by the addition result.
To accomplish the other object of the present invention, there is also provided a motion image decoding apparatus for decoding the output of a motion image encoding apparatus, the motion image decoding apparatus comprising: an object contour restoring unit for decoding received encoded object contour data and restoring object contour data indicating a contour of an object contained in a current image; a differential data decoding unit for decoding the received encoded differential data and outputting differential data; a motion compensation decoding unit for performing a motion compensation operation using the object contour data, prestored reference image data, control points of meshes which divide one of an object contour and a reference image, and motion information involved with the control points; and an addition unit for adding the predictive image data selected by the object contour data and the differential data, and updating the reference image data prestored in the motion compensation decoding unit using the image data obtained by the addition result.
BRIEF DESCRIPTION OF THE DRAWINGS
The preferred embodiments are described with reference to the drawings wherein:
FIG. 1 is a block diagram showing a motion image encoding apparatus according to a preferred embodiment of the present invention;
FIG. 2 is a detailed block diagram of the mesh producer of FIG. 1;
FIG. 3 is a block diagram showing a motion image decoding apparatus for decoding data encoded by the FIG. 1 apparatus;
FIG. 4 is a block diagram showing a motion image encoding apparatus according to another preferred embodiment of the present invention; and
FIG. 5 is a block diagram showing a motion image decoding apparatus for decoding data encoded by the FIG. 4 apparatus.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
Preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.
In FIG. 1 showing a motion image encoding apparatus according to a preferred embodiment of the present invention, an object extractor 10 receives digital image data from an external source, and extracts an arbitrary object from a current image represented by the received image data. Then, the object extractor 10 generates object contour data representing contour information of the extracted object. The object contour data is supplied to a differential data encoder 20 and a decoding apparatus of FIG. 3 to be described later. The object contour data contains spatial positions of pixels which determine a contour of the object or spatial positions of segments composed of the pixels. The object extractor 10 outputs the object contour data and the received current image data to a mesh generator 12. The mesh generator 12 discriminates an object region from the current image based on the object contour data, and performs a signal processing for regular mesh representation with respect to the image in the object region. The detailed structure and operation of the mesh generator 12 will be described with reference to FIG. 2.
A block former 121 in the mesh generator 12 receives the current image data and the object contour data output from the object extractor 10. The block former 121 divides the current image represented by the received image data into image blocks each having a predetermined size, and supplies image block data representing image blocks and the object contour data to a block selector 123. The block selector 123 selects image blocks containing the image data in the object region among the entire image blocks constituting the current image based on the object contour data. The data representing the selected image blocks and the object contour data is output to an object image composer 125. Here, image blocks having only image data which do not belong to the object region, are not selected. Therefore, the data representing the image blocks which are not selected is not supplied to the object image composer 125. The object image composer 125 which receives the data of the selected image blocks merges the selected image blocks and supplies the data representing the object image generated from the merging result and the object contour data to a mesh composer 127. The mesh composer 127 receiving the object image data divides the object image into a regular rectangular mesh or a regular triangular mesh.
In the case that an image is divided using regular rectangular meshes, the mesh composer 127 divides the object image into grids each having a predetermined size different from a block size. Here, a grid has a size smaller than a block. Then, the mesh composer 127 determines vertices contained in the object region among the vertices of the grids obtained by the division as control points, and determines grids having the control points as regular rectangular meshes. The mesh composer 127 outputs rectangular mesh data containing the image data involving the control points of the regular rectangular meshes to a motion estimation and compensation portion 14 of FIG. 1. Here, the image data involving each control point includes the position of a control point and a gray value at the position of the control point, and the positions of the pixels having positions adjacent to the control point and the gray values. The mesh composer 127 also outputs the current image data to a difference value calculator 18.
In the case that an image is divided using regular triangular meshes, the mesh composer 127 divides each regular rectangular mesh obtained through the above process into two regular triangular meshes based on similarity of the image data in the mesh. To judge similarity of the image data of each regular rectangular mesh, the mesh composer 127 compares a difference value between pixel values corresponding to two control points which are located at a diagonal direction of 45° in the regular rectangular mesh, with a difference value between pixel values corresponding to two control points located at a diagonal direction of −45° therein in size. The mesh composer 127 divides the rectangular mesh in the diagonal direction corresponding to the smaller difference value. As a result, two regular triangular meshes are generated for each rectangular mesh. The mesh composer 127 outputs the triangular mesh data containing the image data involving the control points of the regular triangular meshes to the motion estimation and compensation portion 14. The mesh composer 127 outputs the current image data to a difference value calculator 18 as well.
The motion estimation and compensation portion 14 performs motion estimation and compensation using the reference image data stored in the memory 16 and the mesh data involving the object image in the current image received from the mesh former 12. The motion estimation and compensation portion 14 first compares in magnitude the image data involved in the control points of each mesh with the reference image data stored in the memory 16. Each control point ideally corresponds to a single pixel. That is, a control point is expressed by a spatial position of a pixel and a gray value. However, it is actually difficult to exactly find out a position having image data similar to the control point in the reference image by only pixel to pixel comparison. Thus, the image data involving each control point is image data which is involved in a pixel corresponding to the control point and a plurality of pixels neighboring the pixel corresponding to the control point, as described in the mesh composer 127. The motion estimation and compensation portion 14 determines a position in the reference image having the image data which is most similar to the image data of the control point based on the data magnitude comparison result. Then, the motion estimation and compensation portion 14 determines a motion vector representing movement between a control point and a corresponding position in the reference image. If the motion vectors are determined with respect to all the control points, the motion estimation and compensation portion 14 performs spatial transformation such as image warping using the determined motion vectors, and generates a predictive image with respect to the current image. The data representing the predictive image is output to the difference value calculator 18 and the adder 24. The motion information representing the motion vectors corresponding to all the control points is supplied to the FIG. 3 apparatus. Since the above-described spatial transformation is well known in the field of a digital image processing, the detailed description thereof will be omitted.
The difference value calculator 18 calculates difference values between the current image data received from the mesh generator 12 and the predictive image data output from the motion estimation and compensation portion 14. The difference value calculation is performed between the pixels having corresponding positions with respect to the current image and the predictive image, and the resultant difference value data is output to a differential data encoder 20. The differential data encoder 20 encodes difference value data in the object region determined by the object contour data of the object extractor 10 among the difference value data supplied from the difference value calculator 18. If the data encoded by the differential data encoder 20 is defined as difference value data in the object region, only difference value data involving the object of the current image is encoded. Accordingly, the differential data encoder 20 can perform a more efficient encoding operation with respect to the current image. The differential data encoder 20 encodes the object contour data received from the object extractor 10 as well. To encode the difference value data and the object contour data in the object region, the differential data encoder 20 uses a well-known orthogonal transform coding method including discrete cosine transform (DCT) and so on. The encoded object contour data, the encoded difference value data and the motion information are transmitted to a decoding apparatus of FIG. 3 to be described later via a transmission channel (not shown), or are recorded on a storage medium used in the decoding apparatus of FIG. 3.
A differential data decoder 22 receiving the encoded difference value data from the differential data encoder 20 restores the difference value data via a reverse procedure of the signal processing of the differential data encoder 20. The adder 24 adds the predictive image data output from the motion estimation and compensation portion 14 and the difference value data of the differential data decoder 22, and outputs the added result to the memory 16. The data output from the adder 24 is data of the current image from which motion estimation and compensation has been performed, and is stored in the memory 16 to be used as reference image data for motion estimation and motion compensation with respect to a next image.
A motion image decoding apparatus shown in FIG. 3 receives the encoded difference value data, the encoded object contour data and the motion information generated by the FIG. 1 apparatus. An object contour restorer 30 decodes the encoded object contour data. The object contour data is output to a mesh generator 34. The mesh generator 34 receiving the object contour data divides the entire image into grids each having a predetermined size, determines vertices existing in the object region determined by the object contour data as control points among the vertices of the grids obtained by the division, and determines the grids having the control points as rectangular meshes. In the case that the mesh generator 12 of FIG. 1 is designed to generate mesh data representing the rectangular meshes, the mesh generator 34 also generates mesh data representing rectangular meshes. Meanwhile, in the case that the mesh generator 12 of FIG. 1 is designed to generate the mesh data involving triangular meshes, the mesh generator 34 generates mesh data involving the triangular meshes. The mesh generator 34 supplies the generated mesh data to a motion compensator 38 and supplies the object contour data to an adder 36.
Meanwhile, a differential data decoder 32 decodes the encoded difference value data generated by the FIG. 1 apparatus. The difference value data is output to the adder 36. The motion compensator 38 receives the mesh data output from the mesh generator 34 and the motion information generated by the FIG. 1 apparatus. The motion compensator 38 generates predictive image data using control points contained in the mesh data corresponding to a current image, motion vectors corresponding to all the control points and contained in motion information, and the reference image data stored in the memory 40. To generate the predictive image data, the motion compensator 38 uses a motion vector corresponding to each control point to find out a position in the reference image which is most similar to the control point, and then uses spatial transformation such as image warping to generate predictive image data for a current image. The motion compensator 38 outputs the generated predictive image data to an adder 36.
The adder 36 receives the predictive image data output from the motion compensator 38 and the difference value data output from the differential data decoder 32. The adder 36 adds only predictive image data in the object region determined by the object contour data supplied from the mesh generator 34 among the entire predictive image data to the difference value data corresponding thereto. By doing so, image data with respect to an object in the current image can be restored. The output data of the adder 36 is stored in the memory 40 to be used as reference image data for motion compensation of a next image.
FIG. 4 is a block diagram showing a motion image encoding apparatus according to another preferred embodiment of the present invention. Since the blocks of FIG. 4 having the same reference numerals as those of the FIG. 1 blocks perform the same functions as those of the corresponding blocks of FIG. 1, the detailed description thereof will be omitted. The FIG. 1 apparatus uses the meshes obtained from the current image and the reference image data stored in the memory 16 to generate predictive image data. On the contrary, the FIG. 4 apparatus uses meshes obtained from reference image data stored in a memory 48 and externally input current image data to generate predictive image data.
The external current image data is input to the object extractor 10 and a motion estimation and compensation portion 44. The object extractor 10 generates object contour data using the current image data. The object contour data is output to the differential data encoder 20. A mesh generator 42 reads the reference image data stored in the memory 48 and divides the entire reference image into irregular meshes. For such division, the mesh generator 42 divides the reference image into regular rectangular meshes and divides the regular rectangular meshes into regular triangular meshes. Since the method for generation of the regular rectangular meshes and the regular triangular meshes is the same as a signal processing of the FIG. 1 mesh generator 12, the detailed description thereof will be omitted.
Another function of the mesh generator 42 generates an irregular mesh from regular rectangular meshes or regular triangular meshes. The procedure of generating the irregular mesh is as follows. The mesh generator 42 removes one selected among the control points of the regular triangular meshes, and performs triangulation with respect to the region from which the control point is removed. The triangular meshes generated by the triangulation have irregular triangular shapes. Such control point removal and triangulation repeats until the number of the remaining control points is identical to a predetermined value. To remove a particular control point, a quantified difference value is obtained between image descriptiveness obtained when the control point has been removed from a support region of the control point and that when the former has not been removed. Here, the support region is a region surrounded by neighboring control points to a corresponding control point and lines connecting the neighboring control points. A control point corresponding to the smallest difference value among the quantified difference values of the image descriptiveness contributes little to the image descriptiveness, and accordingly can be removed. The irregular mesh generation technique is disclosed in a paper entitled “Irregular Triangular Mesh Representation Based on Adaptive Control Point Removal” published in SPIE's 1996 Symposium on Visual Communications and Image Processing by Kang W. Chun, Byungwoo Jean and Jae M. Jo. The mesh generator 42 outputs information about the finally generated irregular triangular meshes, that is, the irregular mesh data representing the remaining control points and the irregular meshes surrounded by the control points, to a motion estimation and compensation portion 44. The mesh generator 42 outputs the reference image data read from the memory 48 to a difference value calculator 46 as well.
The motion estimation and compensation portion 44 receives the externally supplied current image data and the irregular mesh data of the mesh generator 42, and reads the reference image data stored in the memory 48. The motion estimation and compensation portion 44 determines corresponding positions in the current image having the image data which is most similar to the control points of the irregular meshes. After the corresponding positions are determined, the motion estimation and compensation portion 44 generates motion vectors between the control points and the corresponding positions and uses spatial transformation to generate a predictive image using the current image. The predictive image data is supplied to the difference value calculator 46 and the adder 24, and the motion information representing the motion vectors are transmitted via a transmission channel or recorded on a storage medium to be used for a decoding apparatus.
The difference value calculator 46 generates difference value data between the reference image data received from the mesh generator 42 and the predictive image data from the motion estimation and compensation portion 44, to supply the difference value data to a differential data encoder 20. The differential data encoder 20 encodes the object contour data output from the object extractor 10 and the difference value data output from the difference value calculator 46. The differential data encoder 20 processes the input data in the same manner as those of the corresponding blocks of FIG. 1. Thus, the encoded data output from the differential data encoder 20 includes data obtained by encoding the difference value data in the object region and the encoded object contour data. The encoded data is transmitted to a receiver via a transmission channel (not shown), or is recorded on a recording medium (not shown). The encoded difference value data among the encoded data is transmitted to the differential data decoder 22.
The differential data decoder 22 decodes the difference value data encoded in the differential data encoder 20, to supply the decoded difference value data to the adder 24. The adder 24 adds the predictive image data supplied from the motion estimation and compensation portion 44 and the difference value data of the differential data decoder 22, and the resultant data is stored in the memory 48, to be used as reference image data for motion estimation and motion compensation for a next image.
FIG. 5 shows a motion image decoding apparatus corresponding to the FIG. 4 apparatus. Since blocks having the same reference numerals as those of the FIG. 3 blocks among the blocks shown in FIG. 5 have the same functions as those of the FIG. 3 blocks, the detailed description thereof will be omitted. The encoded object contour data is input to an object contour restorer 30 and the motion information is input to a motion compensator 52. Also, the encoded difference value data is supplied to the differential data decoder 32. The differential data decoder 32 decodes the encoded difference value data to output the decoded difference value data to an adder 56. The object contour restorer 30 decodes the encoded object contour data to output the decoded object contour data to a mesh former 51. The mesh former 51 reads the reference image data stored in a memory 54, and divides the object region in the reference image determined by the object contour data via the same signal processing as that of the FIG. 4 mesh generator 42, into irregular meshes. The mesh data output from the mesh former 51 is supplied to the motion compensator 52. The motion compensator 52 uses the mesh data, the motion information and the reference image data stored in the memory 54 to generate the predictive image data. The adder 56 adds the predictive image data and the difference value data output from the differential data decoder 32. The resultant current image is stored in the memory 54 to be used as reference image data for motion compensation of a next image.
As described above, the encoding and decoding system according to the present invention performs motion estimation and motion compensation based on mesh representation with respect to an arbitrary object contained in an image. As a result, a more efficient encoding can be performed compared with a case when motion estimation and motion compensation is performed with respect to the entire image. Thus, an apparatus appropriate for applications requiring a very low bit rate can be designed.
While only certain embodiments of the invention have been specifically described herein, it will apparent that numerous modifications may be made thereto without departing from the spirit and scope of the invention.

Claims (2)

What is claimed is:
1. An object-oriented image encoding apparatus for performing a motion prediction and compensation in a unit of an object image, the object-oriented image encoding apparatus comprising:
object contour information extraction means for extracting information representing a contour of an object image from an image represented by received image data and outputting the extracted contour information:
mesh production means for representing the object image as a triangular or a rectangular mesh;
motion vector producing means to produce a motion vector on a node point of the mesh in the object image represented as the mesh; and
transmission means for transmitting said contour information, mesh information and said motion vector on the node point of the mesh.
2. An object-oriented image encoding method for performing a motion prediction and compensation in a unit of an object image, the object-oriented image encoding method comprising the steps of:
(a) extracting information representing a contour of an object image from an image represented by received image data and outputting the extracted contour information:
(b) representing the object image as a triangular or a rectangular mesh;
(c) producing a motion vector on a node point of the mesh; and
(d) transmitting said contour information, mesh information and said motion vector on the node point of the mesh.
US09/504,934 1996-05-29 2000-02-16 Encoding and decoding system of motion image containing arbitrary object Expired - Lifetime US6236680B1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
US09/504,934 US6236680B1 (en) 1996-05-29 2000-02-16 Encoding and decoding system of motion image containing arbitrary object
US09/842,405 US6744817B2 (en) 1996-05-29 2001-04-26 Motion predictive arbitrary visual object encoding and decoding system

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
KR1019960018612A KR100215451B1 (en) 1996-05-29 1996-05-29 System for encoding and decoding moving image including objects of any shape
KR96-18612 1996-05-29
US08/864,992 US6038258A (en) 1996-05-29 1997-05-29 Encoding and decoding system of motion image containing arbitrary object
US09/504,934 US6236680B1 (en) 1996-05-29 2000-02-16 Encoding and decoding system of motion image containing arbitrary object

Related Parent Applications (1)

Application Number Title Priority Date Filing Date
US08/864,992 Continuation US6038258A (en) 1996-05-29 1997-05-29 Encoding and decoding system of motion image containing arbitrary object

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US09/842,405 Continuation US6744817B2 (en) 1996-05-29 2001-04-26 Motion predictive arbitrary visual object encoding and decoding system

Publications (1)

Publication Number Publication Date
US6236680B1 true US6236680B1 (en) 2001-05-22

Family

ID=37402254

Family Applications (3)

Application Number Title Priority Date Filing Date
US08/864,992 Expired - Lifetime US6038258A (en) 1996-05-29 1997-05-29 Encoding and decoding system of motion image containing arbitrary object
US09/504,934 Expired - Lifetime US6236680B1 (en) 1996-05-29 2000-02-16 Encoding and decoding system of motion image containing arbitrary object
US09/842,405 Expired - Fee Related US6744817B2 (en) 1996-05-29 2001-04-26 Motion predictive arbitrary visual object encoding and decoding system

Family Applications Before (1)

Application Number Title Priority Date Filing Date
US08/864,992 Expired - Lifetime US6038258A (en) 1996-05-29 1997-05-29 Encoding and decoding system of motion image containing arbitrary object

Family Applications After (1)

Application Number Title Priority Date Filing Date
US09/842,405 Expired - Fee Related US6744817B2 (en) 1996-05-29 2001-04-26 Motion predictive arbitrary visual object encoding and decoding system

Country Status (7)

Country Link
US (3) US6038258A (en)
EP (3) EP1715695A3 (en)
JP (1) JPH1056643A (en)
KR (1) KR100215451B1 (en)
CN (2) CN1134988C (en)
DE (1) DE69736852T2 (en)
ES (1) ES2277350T3 (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030063672A1 (en) * 1999-12-09 2003-04-03 Nathalie Laurent-Chatenet Method for estimating the motion between two digital images with management of mesh overturning and corresponding coding method
US6744817B2 (en) * 1996-05-29 2004-06-01 Samsung Electronics Co., Ltd. Motion predictive arbitrary visual object encoding and decoding system
US20040240543A1 (en) * 2001-09-04 2004-12-02 Faroudja Yves C. Low bandwidth video compression
US20060115001A1 (en) * 2002-04-29 2006-06-01 Nokia Corporation Indicating regions within a picture
US7095786B1 (en) 2003-01-11 2006-08-22 Neo Magic Corp. Object tracking using adaptive block-size matching along object boundary and frame-skipping when object motion is low

Families Citing this family (24)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100535630B1 (en) * 1998-06-30 2006-06-23 주식회사 팬택앤큐리텔 Encoding / Decoding Method of Digital Gray Shape Information / Color Information
KR100620715B1 (en) * 1998-06-30 2007-04-25 주식회사 팬택앤큐리텔 Encoding / Decoding Method of Digital Gray Shape Information / Color Information
WO2000025232A1 (en) * 1998-10-23 2000-05-04 Siemens Aktiengesellschaft Method and array for coding, decoding, and transmitting a digitized image
US6351267B1 (en) * 1998-12-10 2002-02-26 Gizmoz Ltd Fast transmission of graphic objects
JP4126126B2 (en) * 1998-12-11 2008-07-30 株式会社日立製作所 Transmission system and transmission method
JP3897476B2 (en) * 1999-02-15 2007-03-22 キヤノン株式会社 Image processing apparatus and method, and computer-readable memory
KR100611999B1 (en) 1999-08-27 2006-08-11 삼성전자주식회사 Motion compensating method in object based quad-tree mesh using greedy algorithm
US6909746B2 (en) 2001-03-30 2005-06-21 Koninklijke Philips Electronics N.V. Fast robust data compression method and system
CN1838775B (en) * 2001-11-30 2011-11-23 株式会社Ntt都科摩 Moving picture coding apparatus, moving picture decoding apparatus, moving picture coding method and moving picture decoding method
US7373353B2 (en) 2002-05-10 2008-05-13 International Business Machines Corporation Reducing index size for multi-level grid indexes
US7383275B2 (en) 2002-05-10 2008-06-03 International Business Machines Corporation Methods to improve indexing of multidimensional databases
US7143098B2 (en) * 2002-05-10 2006-11-28 International Business Machines Corporation Systems, methods, and computer program products to reduce computer processing in grid cell size determination for indexing of multidimensional databases
JP2005184626A (en) * 2003-12-22 2005-07-07 Canon Inc Image processing apparatus
US20050198008A1 (en) * 2004-03-02 2005-09-08 Adler David W. Index exploitation for spatial data
US7389283B2 (en) * 2004-12-07 2008-06-17 International Business Machines Corporation Method for determining an optimal grid index specification for multidimensional data
CN103118252B (en) 2005-09-26 2016-12-07 三菱电机株式会社 Dynamic image encoding device and dynamic image decoding device
EP1966756B1 (en) * 2005-12-19 2009-07-22 Koninklijke Philips Electronics N.V. Method for facilitating post-processing of images using deformable meshes
KR20080107965A (en) * 2007-06-08 2008-12-11 삼성전자주식회사 Method and apparatus for encoding and decoding image using object boundary based partition
JP5230372B2 (en) * 2008-11-25 2013-07-10 キヤノン株式会社 Image processing apparatus and image processing method
JP5661359B2 (en) * 2010-07-16 2015-01-28 キヤノン株式会社 Image processing apparatus, image processing method, and program
JP5558949B2 (en) * 2010-07-16 2014-07-23 キヤノン株式会社 Image processing apparatus, image processing method, and program
US9189884B2 (en) * 2012-11-13 2015-11-17 Google Inc. Using video to encode assets for swivel/360-degree spinners
CN106331722B (en) 2015-07-03 2019-04-26 华为技术有限公司 Image prediction method and relevant device
CN111526361B (en) 2016-02-06 2022-05-13 华为技术有限公司 Image coding and decoding method and device

Citations (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5274453A (en) 1990-09-03 1993-12-28 Canon Kabushiki Kaisha Image processing system
JPH07170523A (en) 1993-03-04 1995-07-04 Toshiba Corp Moving picture coder and moving picture decoder
US5598215A (en) 1993-05-21 1997-01-28 Nippon Telegraph And Telephone Corporation Moving image encoder and decoder using contour extraction
US5635929A (en) 1995-02-13 1997-06-03 Hughes Aircraft Company Low bit rate video encoder and decoder
US5654771A (en) 1995-05-23 1997-08-05 The University Of Rochester Video compression system using a dense motion vector field and a triangular patch mesh overlay model
US5689306A (en) 1995-01-09 1997-11-18 Daewoo Electronics Co., Ltd. Method and apparatus for encoding a video signal using pixel-by-pixel motion prediction
US5719629A (en) 1995-12-27 1998-02-17 Samsung Electronics Co., Ltd. Motion picture encoding method and apparatus thereof
US5736987A (en) 1996-03-19 1998-04-07 Microsoft Corporation Compression of graphic data normals
US5751365A (en) 1995-08-04 1998-05-12 Nec Corporation Motion compensated inter-frame prediction method and apparatus using motion vector interpolation with adaptive representation point addition
US5815601A (en) * 1995-03-10 1998-09-29 Sharp Kabushiki Kaisha Image encoder and image decoder
US5815685A (en) * 1994-09-16 1998-09-29 Mitsubishi Denki Kabushiki Kaisha Apparatus and method for correcting light proximity effects by predicting mask performance
US5818463A (en) 1997-02-13 1998-10-06 Rockwell Science Center, Inc. Data compression for animated three dimensional objects
US5819070A (en) * 1995-04-27 1998-10-06 Canon Kabushiki Kaisha Method and apparatus for processing finite element meshing model
US5907626A (en) 1996-08-02 1999-05-25 Eastman Kodak Company Method for object tracking and mosaicing in an image sequence using a two-dimensional mesh
US5930494A (en) * 1995-06-28 1999-07-27 Nec Corporation Process simulator and process simulating method
US5936671A (en) 1996-07-02 1999-08-10 Sharp Laboratories Of America, Inc. Object-based video processing using forward-tracking 2-D mesh layers
US6038258A (en) * 1996-05-29 2000-03-14 Samsung Electronics Co., Ltd. Encoding and decoding system of motion image containing arbitrary object
US6047088A (en) * 1996-12-16 2000-04-04 Sharp Laboratories Of America, Inc. 2D mesh geometry and motion vector compression
US6058212A (en) * 1996-01-17 2000-05-02 Nec Corporation Motion compensated interframe prediction method based on adaptive motion vector interpolation

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR0159575B1 (en) * 1994-10-31 1999-01-15 배순훈 Discrete processing apparatus between neighbor region in region based coding
KR100235343B1 (en) * 1994-12-29 1999-12-15 전주범 Apparatus for calculating motion vector in encoder using segmentation method
KR0159370B1 (en) * 1995-03-20 1999-01-15 배순훈 Method and apparatus for encoding a video signals using a boundary of an object
US5646867A (en) * 1995-07-24 1997-07-08 Motorola Inc. Method and system for improved motion compensation
US5691769A (en) * 1995-09-07 1997-11-25 Daewoo Electronics Co, Ltd. Apparatus for encoding a contour of an object
DE69724899T2 (en) * 1996-05-06 2004-07-08 Koninklijke Philips Electronics N.V. METHOD AND SYSTEM FOR CODING AND DECODING SEGMENTED VIDEO DATA
US6148026A (en) * 1997-01-08 2000-11-14 At&T Corp. Mesh node coding to enable object based functionalities within a motion compensated transform video coder

Patent Citations (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5274453A (en) 1990-09-03 1993-12-28 Canon Kabushiki Kaisha Image processing system
JPH07170523A (en) 1993-03-04 1995-07-04 Toshiba Corp Moving picture coder and moving picture decoder
US5598215A (en) 1993-05-21 1997-01-28 Nippon Telegraph And Telephone Corporation Moving image encoder and decoder using contour extraction
US5815685A (en) * 1994-09-16 1998-09-29 Mitsubishi Denki Kabushiki Kaisha Apparatus and method for correcting light proximity effects by predicting mask performance
US5689306A (en) 1995-01-09 1997-11-18 Daewoo Electronics Co., Ltd. Method and apparatus for encoding a video signal using pixel-by-pixel motion prediction
US5635929A (en) 1995-02-13 1997-06-03 Hughes Aircraft Company Low bit rate video encoder and decoder
US5978515A (en) * 1995-03-10 1999-11-02 Sharp Kabushiki Kaisha Image encoder and image decoder
US5815601A (en) * 1995-03-10 1998-09-29 Sharp Kabushiki Kaisha Image encoder and image decoder
US5819070A (en) * 1995-04-27 1998-10-06 Canon Kabushiki Kaisha Method and apparatus for processing finite element meshing model
US5654771A (en) 1995-05-23 1997-08-05 The University Of Rochester Video compression system using a dense motion vector field and a triangular patch mesh overlay model
US5930494A (en) * 1995-06-28 1999-07-27 Nec Corporation Process simulator and process simulating method
US5751365A (en) 1995-08-04 1998-05-12 Nec Corporation Motion compensated inter-frame prediction method and apparatus using motion vector interpolation with adaptive representation point addition
US5719629A (en) 1995-12-27 1998-02-17 Samsung Electronics Co., Ltd. Motion picture encoding method and apparatus thereof
US6058212A (en) * 1996-01-17 2000-05-02 Nec Corporation Motion compensated interframe prediction method based on adaptive motion vector interpolation
US5736987A (en) 1996-03-19 1998-04-07 Microsoft Corporation Compression of graphic data normals
US6038258A (en) * 1996-05-29 2000-03-14 Samsung Electronics Co., Ltd. Encoding and decoding system of motion image containing arbitrary object
US5936671A (en) 1996-07-02 1999-08-10 Sharp Laboratories Of America, Inc. Object-based video processing using forward-tracking 2-D mesh layers
US5907626A (en) 1996-08-02 1999-05-25 Eastman Kodak Company Method for object tracking and mosaicing in an image sequence using a two-dimensional mesh
US6047088A (en) * 1996-12-16 2000-04-04 Sharp Laboratories Of America, Inc. 2D mesh geometry and motion vector compression
US5818463A (en) 1997-02-13 1998-10-06 Rockwell Science Center, Inc. Data compression for animated three dimensional objects

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
Image Communication, vol. 2, No. 4, dated Dec. 1990, pp. 409-428.

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6744817B2 (en) * 1996-05-29 2004-06-01 Samsung Electronics Co., Ltd. Motion predictive arbitrary visual object encoding and decoding system
US20030063672A1 (en) * 1999-12-09 2003-04-03 Nathalie Laurent-Chatenet Method for estimating the motion between two digital images with management of mesh overturning and corresponding coding method
US6963605B2 (en) * 1999-12-09 2005-11-08 France Telecom (Sa) Method for estimating the motion between two digital images with management of mesh overturning and corresponding coding method
US20040240543A1 (en) * 2001-09-04 2004-12-02 Faroudja Yves C. Low bandwidth video compression
US20060115001A1 (en) * 2002-04-29 2006-06-01 Nokia Corporation Indicating regions within a picture
US20070110154A1 (en) * 2002-04-29 2007-05-17 Nokia Corporation Random access points in video encoding
US7760802B2 (en) 2002-04-29 2010-07-20 Nokia Corporation Random access points in video encoding
US7826531B2 (en) * 2002-04-29 2010-11-02 Nokia Corporation Indicating regions within a picture
US7095786B1 (en) 2003-01-11 2006-08-22 Neo Magic Corp. Object tracking using adaptive block-size matching along object boundary and frame-skipping when object motion is low
US7142600B1 (en) 2003-01-11 2006-11-28 Neomagic Corp. Occlusion/disocclusion detection using K-means clustering near object boundary with comparison of average motion of clusters to object and background motions
USRE42790E1 (en) 2003-01-11 2011-10-04 Neomagic Corporation Occlusion/disocclusion detection using K-means clustering near object boundary with comparison of average motion of clusters to object and background motions

Also Published As

Publication number Publication date
US20010014119A1 (en) 2001-08-16
CN1254238A (en) 2000-05-24
EP1523196A2 (en) 2005-04-13
CN1146242C (en) 2004-04-14
DE69736852D1 (en) 2006-12-07
EP1715695A2 (en) 2006-10-25
ES2277350T3 (en) 2007-07-01
EP0810792B1 (en) 2006-10-25
US6038258A (en) 2000-03-14
EP1523196A3 (en) 2006-01-18
EP0810792A2 (en) 1997-12-03
CN1177259A (en) 1998-03-25
JPH1056643A (en) 1998-02-24
EP1715695A3 (en) 2008-10-22
DE69736852T2 (en) 2007-09-06
KR100215451B1 (en) 1999-08-16
CN1134988C (en) 2004-01-14
KR970078652A (en) 1997-12-12
US6744817B2 (en) 2004-06-01
EP0810792A3 (en) 2000-12-13

Similar Documents

Publication Publication Date Title
US6236680B1 (en) Encoding and decoding system of motion image containing arbitrary object
US5086439A (en) Encoding/decoding system utilizing local properties
US5592228A (en) Video encoder using global motion estimation and polygonal patch motion estimation
JP3277111B2 (en) Video encoding device and video decoding device
US8599923B2 (en) Motion vector encoding device and decoding device
EP1379000B1 (en) Signal encoding method and apparatus and decoding method and apparatus
CN101682775B (en) Motion vector searching method and device
JP4993676B2 (en) Image coding apparatus and image coding method
JP2009267710A (en) Image prediction encoding device, image prediction encoding method, image prediction encoding program, image prediction decoding device, image prediction decoding method, and image prediction decoding program
EP2161937A1 (en) Video encoding method, decoding method, device thereof, program thereof, and recording medium containing the program
US6020923A (en) Method and apparatus for coding and recording an image signal and recording medium for storing an image signal
WO1995004432A1 (en) Coding image data
JPH08317410A (en) Method and apparatus for coding image signal
KR19980018127A (en) A picture coding method, a picture decoding method, a picture coding apparatus using the method, a picture decoding apparatus and a recording medium for recording the method
CA2087876C (en) Moving-image signal encoding apparatus
JP3226637B2 (en) Coding device and decoding device for predictive coding method
JPH0537916A (en) Image transmission equipment
EP0929976B1 (en) Method and apparatus for motion estimation
JPH08214169A (en) Fractal image coding system
KR100240344B1 (en) Adaptive vertex coding apparatus and method
KR100207391B1 (en) Image coding system and moving information detecting method using adaptive vector quantization
JP2005236723A (en) Device and method for encoding moving image, and device and method for decoding the moving image
KR0178203B1 (en) The improved pattern vector coding system
JPH06225288A (en) Movement compensation predictive coder/decoder
Saito et al. Some variants of universal pattern-matching interframe coding

Legal Events

Date Code Title Description
STCF Information on status: patent grant

Free format text: PATENTED CASE

FEPP Fee payment procedure

Free format text: PAYOR NUMBER ASSIGNED (ORIGINAL EVENT CODE: ASPN); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY

FPAY Fee payment

Year of fee payment: 4

FPAY Fee payment

Year of fee payment: 8

FEPP Fee payment procedure

Free format text: PAYER NUMBER DE-ASSIGNED (ORIGINAL EVENT CODE: RMPN); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY

Free format text: PAYOR NUMBER ASSIGNED (ORIGINAL EVENT CODE: ASPN); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY

FPAY Fee payment

Year of fee payment: 12